Unfortunately, this does not answer the question. In the case of overwhelmed hospitals, the ratio (deaths / serious cases) should be bounded from below by a con

Unfortunately, this does not answer the question. In the case of overwhelmed hospitals, the ratio (deaths / serious cases) should be bounded from below by a constant close to 1. The number of serious cases is the only useful datapoint, but this gives a very bad lower bound on the question, even if the age distribution is known. This data is similar to what Worldometer has, and that is just a mix of confused and misleading information.